Smart Digital Signage with mmWave Radar & Real-Time Content Logic
Budget: ₹5,000 – ₹15,000 INR
I am looking for an IoT / Full-Stack Developer to build the core software for an industrial-grade digital ad network. The system will run on Radxa Cubie A5E (8-Core) hardware and use a mmWave presence radar to trigger advertisements based on real-time pedestrian detection.
The application must be a "State-Aware" media controller that switches between informational content and high-value ads based on audience presence.
Key Features & Logic:
Idle State: When no one is detected, the screen must play a Brand Loop (Video) with a Live News Ticker (RSS Feed) overlaid at the bottom.
Trigger State: Upon radar detection (UART/Serial), the system must immediately interrupt the idle loop and play a Full-Screen Client Ad without lag.
Analytics: The system must log every "Impression" (Ad ID, Start Time, End Time, and Dwell Time) to a local SQLite database.
Cloud Sync: Periodically sync local logs to a central server and "Pull" new ad schedules/video files from a remote CMS.
Technical Stack Requirements:
Language: Python (utilizing Multiprocessing/Threading for zero-lag playback).
Video Engine: mpv (with hardware acceleration) or VLC (via Python bindings).
Hardware Interfacing: pyserial for communication with DFRobot C4001 or HLK-LD2450 sensors.
Database: SQLite (Edge) and PostgreSQL (Server).
Protocols: MQTT for real-time heartbeats and REST API for file syncing.
Deliverables:
Edge Software: A robust, auto-booting Python application for the Radxa A5E.
Server Backend: A simple API to receive logs and serve a JSON manifest of ad schedules.
Basic Admin Dashboard: A web-based view to monitor device "Heartbeats" (Online/Offline status) and view footfall charts.
What I need built is the production-grade cloud layer and its companion web dashboard.
Core platform requirements
• Microservice-oriented or similarly modular architecture that can scale out horizontally as new venues come online.
• Low-friction onboarding for additional edge devices (secure registration, token refresh, OTA update support).
• Event pipeline that ingests real-time trigger data from devices, records impressions, and hands off the correct creative with sub-second latency.
• High availability and fault tolerance baked in (container orchestration such as Kubernetes/EKS/GKE is acceptable; serverless patterns are fine if they still allow persistent WebSocket/MQTT connections).
Dashboard must expose
• Real-time analytics: impressions, dwell-time estimates, trigger frequency, device bandwidth usage.
• Ad scheduling that lets a Content Manager upload 4K creatives, tag them to campaigns, and set time/day or rule-based playback windows.
• Device health monitoring: online/offline status, CPU temp, storage, radar heartbeat, and remote reboot/update controls.
User roles
• Administrator – full system control, user management, deployment parameters.
• Content Manager – campaign creation, creative upload, schedule edits, analytics read-only.
Security & compliance
JWT or OAuth2 authentication, TLS everywhere, audit log retention, and role-based access must be implemented from day one. GDPR-friendly data handling is expected; no personal data is captured, only anonymous counts.
Preferred tech stack
Open to Node.js/TypeScript, Go, or Python for the services; React or Vue for the dashboard; PostgreSQL or Timescale for structured data; Redis/Kafka for queues and pub-sub. Feel free to suggest an alternative if it maintains the scalability priority.
Deliverables
1. Infrastructure-as-code templates (Terraform, Pulumi, or similar) that spin up the entire backend.
2. Source for all services, documented and unit-tested.
3. Responsive web dashboard matching the feature list.
4. API & schema documentation.
5. Deployment guide plus a brief hand-off session.
Acceptance criteria will be a working demo with three simulated edge devices triggering ads, live metrics visible on the dashboard, and the ability to push a new campaign without downtime.
If this aligns with your expertise in distributed systems, edge-to-cloud message flows, and rich admin UIs, let’s set up a short technical chat to confirm approach and timeline.
The application must be a "State-Aware" media controller that switches between informational content and high-value ads based on audience presence.
Key Features & Logic:
Idle State: When no one is detected, the screen must play a Brand Loop (Video) with a Live News Ticker (RSS Feed) overlaid at the bottom.
Trigger State: Upon radar detection (UART/Serial), the system must immediately interrupt the idle loop and play a Full-Screen Client Ad without lag.
Analytics: The system must log every "Impression" (Ad ID, Start Time, End Time, and Dwell Time) to a local SQLite database.
Cloud Sync: Periodically sync local logs to a central server and "Pull" new ad schedules/video files from a remote CMS.
Technical Stack Requirements:
Language: Python (utilizing Multiprocessing/Threading for zero-lag playback).
Video Engine: mpv (with hardware acceleration) or VLC (via Python bindings).
Hardware Interfacing: pyserial for communication with DFRobot C4001 or HLK-LD2450 sensors.
Database: SQLite (Edge) and PostgreSQL (Server).
Protocols: MQTT for real-time heartbeats and REST API for file syncing.
Deliverables:
Edge Software: A robust, auto-booting Python application for the Radxa A5E.
Server Backend: A simple API to receive logs and serve a JSON manifest of ad schedules.
Basic Admin Dashboard: A web-based view to monitor device "Heartbeats" (Online/Offline status) and view footfall charts.
What I need built is the production-grade cloud layer and its companion web dashboard.
Core platform requirements
• Microservice-oriented or similarly modular architecture that can scale out horizontally as new venues come online.
• Low-friction onboarding for additional edge devices (secure registration, token refresh, OTA update support).
• Event pipeline that ingests real-time trigger data from devices, records impressions, and hands off the correct creative with sub-second latency.
• High availability and fault tolerance baked in (container orchestration such as Kubernetes/EKS/GKE is acceptable; serverless patterns are fine if they still allow persistent WebSocket/MQTT connections).
Dashboard must expose
• Real-time analytics: impressions, dwell-time estimates, trigger frequency, device bandwidth usage.
• Ad scheduling that lets a Content Manager upload 4K creatives, tag them to campaigns, and set time/day or rule-based playback windows.
• Device health monitoring: online/offline status, CPU temp, storage, radar heartbeat, and remote reboot/update controls.
User roles
• Administrator – full system control, user management, deployment parameters.
• Content Manager – campaign creation, creative upload, schedule edits, analytics read-only.
Security & compliance
JWT or OAuth2 authentication, TLS everywhere, audit log retention, and role-based access must be implemented from day one. GDPR-friendly data handling is expected; no personal data is captured, only anonymous counts.
Preferred tech stack
Open to Node.js/TypeScript, Go, or Python for the services; React or Vue for the dashboard; PostgreSQL or Timescale for structured data; Redis/Kafka for queues and pub-sub. Feel free to suggest an alternative if it maintains the scalability priority.
Deliverables
1. Infrastructure-as-code templates (Terraform, Pulumi, or similar) that spin up the entire backend.
2. Source for all services, documented and unit-tested.
3. Responsive web dashboard matching the feature list.
4. API & schema documentation.
5. Deployment guide plus a brief hand-off session.
Acceptance criteria will be a working demo with three simulated edge devices triggering ads, live metrics visible on the dashboard, and the ability to push a new campaign without downtime.
If this aligns with your expertise in distributed systems, edge-to-cloud message flows, and rich admin UIs, let’s set up a short technical chat to confirm approach and timeline.
Related categories:
Python
Node.js
PostgreSQL
Ubuntu
Redis
React.js
MQTT
RESTful API
Containerization